Methane (CH₄) is a potent greenhouse gas, and its reduction is crucial for limiting global warming to 1.5 °C. The Global Methane Pledge aims to reduce anthropogenic CH₄ emissions by 30% by 2030. Effective mitigation strategies require a comprehensive understanding of emission sources and their spatiotemporal variability. Current CH₄ emission studies utilize bottom-up and top-down approaches. Bottom-up methods analyze individual source sectors, but discrepancies often exist between total emission estimates and atmospheric observations due to uncertainties in emission factors and activity data. Top-down methods use inversion techniques to constrain total emissions using atmospheric observations, but lack the granularity to distinguish individual sources. Isotopic composition (δ¹³C-CH₄) provides insights into CH₄ origins, as different sources exhibit characteristic isotopic signatures. Previous studies incorporating δ¹³C-CH₄ measurements have yielded diverse conclusions about the drivers of CH₄ growth rate changes, with some emphasizing microbial sources and others highlighting competing contributions from both microbial and fossil fuel sources. This study addresses these discrepancies by employing a forward modeling approach using a 3D atmospheric chemistry transport model (MIROC4-ACTM) that simulates the history of CH₄, δ¹³C-CH₄ and δD-CH₄ from 1970 to 2020. The simulations integrate various emission inventories, isotopic source signatures, and chemical sink parameters, allowing assessment of different emission scenarios and their consistency with atmospheric observations. The comparison of simulated and observed atmospheric trends across latitudinal bands and vertical profiles helps to identify the most plausible changes in sector-specific emissions that are needed to reconcile the global CH₄ budget.

Numerous studies have employed bottom-up and top-down approaches to estimate global methane emissions. Bottom-up studies provide detailed sector-specific emission estimates but often suffer from uncertainties in emission factors and activity data, leading to inconsistencies with atmospheric observations. Top-down methods, using atmospheric observations and inversion techniques, provide constraints on total emissions but lack the resolution to identify individual source contributions. The use of isotopic composition (δ¹³C-CH₄) has enhanced source apportionment, as different sources possess distinct isotopic signatures. However, even studies incorporating δ¹³C-CH₄ data have produced varying conclusions regarding the relative contributions of fossil fuel and microbial sources to recent atmospheric methane growth. Some studies emphasize increased emissions from microbial sources, while others suggest competing contributions from both microbial and fossil fuel sources. The discrepancies stem from the inherent uncertainties associated with bottom-up inventories, especially for fugitive fossil fuel emissions (oil and gas, coal), the uncertainty in the isotopic source signatures, and the complexities of atmospheric transport and chemical processes. Incorporation of the isotopic data in simplified box models or complex 3D inversions also leads to conflicting results.

This study utilizes the MIROC (version 4)-based atmospheric chemistry-transport model (MIROC4-ACTM) for forward simulations of atmospheric CH₄, δ¹³C-CH₄, and δD-CH₄ from 1970 to 2020. The model incorporates various data sources, including emission inventories (EDGARv6, GAINSv4), kinetic isotope effect (KIE) values for different chemical sinks (OH, Cl, O('D)), chlorine fields, and region-specific source signatures. Several sensitivity simulations were conducted using different combinations of emission estimates from various inventories (EDGARv6, GAINSv4) and isotopic source signatures. These simulations were evaluated against extensive balloon-based vertical measurements, and surface observations covering multiple latitudes and decades. Rigorous testing involved verifying the choice of initial atmospheric CH₄ and δ¹³C-CH₄ values, monitoring tracer mass conservation, and evaluating the impact of various atmospheric chemistry parameters such as tropospheric Cl sink and KIEOH values. The simulations were compared to observations using long-term trends, growth rates, latitudinal gradients, and vertical profiles. The goal was to identify sector-specific emission changes that best align with observations of CH₄ and δ¹³C-CH₄ across various spatial and temporal scales. This involved systematic adjustments to fugitive fossil fuel emissions (ONG, coal), focusing on the discrepancies between different inventories and atmospheric observations. In addition, sensitivity analyses were conducted by evaluating the influence of isotope parameters (KIE values and source signatures) on simulated δ¹³C-CH₄.

The study's key findings challenge the consistent increase in microbial and fossil fuel emissions reported by the EDGARv6 inventory. The analysis demonstrates that the EDGARv6 inventory significantly overestimates the observed CH₄ growth rate and fails to reproduce the observed δ¹³C-CH₄ trends. Conversely, using the GAINSv4 inventory (after excluding unconventional emissions from the USA after 2006), the model successfully replicates observed trends and latitudinal/vertical distributions of atmospheric CH₄ and δ¹³C-CH₄. The decline in ONG emissions from 1990 to 2010 and subsequent stabilization of fossil fuel emissions (offsetting coal increases in China) are found to be crucial to the observed stable CH₄ trend from 2000 to 2020. Meanwhile, microbial emissions (from farming, waste, etc.) have significantly increased. The model's accuracy in representing atmospheric chemistry and transport processes was validated by comparing simulated and observed vertical profiles of CH₄, δ¹³C-CH₄, and δD-CH₄. The sensitivity analysis highlights the importance of accurate KIE values and geographically varying δ¹³C-CH₄ source signatures in correctly simulating the isotopic composition of atmospheric methane and the latitudinal gradient of δ¹³C-CH₄. Specifically, accounting for geographic variability in wetland, biomass burning, and coal signatures significantly improves the model's representation of the seasonal cycle and interannual variability of δ¹³C-CH₄. Adjusting ONG and geological signatures also improves the agreement between simulated and observed data. The study's results suggest that the most plausible scenario includes a decrease in fossil fuel emissions (primarily ONG) between the 1990s and 2010s, coupled with an increase in microbial and biomass burning emissions. The total fossil fuel emission trend in this scenario closely matches the independent inversion estimates.

This study's findings demonstrate that the observed changes in atmospheric CH₄ concentration and its isotopic composition during 1990-2020 cannot be explained solely by a consistent increase in both microbial and fossil fuel emissions as suggested by some inventories. The results highlight the limitations of relying solely on bottom-up estimates and underscore the importance of incorporating atmospheric observations and isotopic constraints. The significant role of reduced ONG emissions in stabilizing atmospheric CH₄ growth in the late 1990s and early 2000s and the increase in microbial and biomass burning emissions as primary drivers of the subsequent CH₄ growth challenges the widely accepted narrative of a continuous rise in fossil fuel emissions. The study further refines the understanding of the global CH₄ budget by quantifying the impact of uncertainties in various components (KIE values, source signatures, atmospheric chemistry, emission inventories) on the overall budget and associated trends. This provides crucial information for developing and evaluating future mitigation strategies and underscores the significance of robust, observation-constrained approaches when interpreting complex atmospheric changes. The results strongly advocate for continued improvements in emission inventories and a greater emphasis on the use of atmospheric observations and isotopic constraints in the assessment of the global CH₄ budget.

This study provides a refined understanding of the drivers of atmospheric CH₄ changes during 1990-2020. It demonstrates that the combination of decreased ONG emissions and increased microbial emissions can explain the observed trends and distributions. The findings emphasize the importance of using multiple data sources (inventories, atmospheric observations, isotopic data) and advanced modeling techniques to comprehensively assess CH₄ emissions and their impact on climate change. Future research should focus on further reducing uncertainties in individual components of the CH₄ budget, including improving the spatial resolution of emission inventories, better characterizing isotopic signatures and atmospheric chemical processes, and refining inversion techniques.

The study's analysis is constrained by the uncertainties inherent in several aspects of the CH₄ budget, including emission inventories (especially for fugitive fossil fuel emissions), isotopic source signatures, KIE values for different chemical sinks, and the representation of atmospheric transport and chemical processes within the model. The assumed constant atmospheric OH sink could also impact the accuracy of the overall assessment. While the study addresses some of these uncertainties through sensitivity analysis, complete exploration of all possible uncertainties is computationally infeasible given the large number of uncertain parameters. Additionally, the model's resolution and representation of specific processes may limit the accuracy of certain results, such as the spatial distribution of CH₄ emissions.